Hugging Face Content Policy

This document is Hugging Face's Content Policy, effective April 10, 2025, incorporated by reference into the company's Terms of Service as a binding agreement, governing all content posted, displayed, or accessed on the Hugging Face Hub platform including ML Artifacts (models, datasets, spaces) and Community Content (discussions, comments, READMEs, model cards). The policy states that users consent to its terms by accessing or using the platform, authorizes the Hugging Face Team to take enforcement actions including unranking, adding audience restriction tags, removing or disabling content, restricting interactions, and suspending or terminating accounts, and establishes a DMCA-compliant takedown and counter-notification process for intellectual property disputes. Notably, the policy explicitly states that content feed display is non-personalized and chronologically ordered without algorithmic curation, which is operationally distinct from many platform content policies; additionally, the policy reserves broad discretionary authority to moderate content beyond its enumerated restricted categories 'in response to evolving challenges posed by advancements in Machine Learning,' without specifying procedural constraints on that discretion. The policy references the EU Digital Services Act (DSA) by name, providing EU residents with access to certified out-of-court dispute settlement bodies and designating legal@huggingface.co as the DSA point of contact for EU authorities; given Hugging Face's operation of an AI-focused hosting platform, the policy also engages considerations under the EU AI Act and applicable national content regulation frameworks. Compliance teams should note that the policy's effective-date update mechanism (10 days notice via website posting, continued use constituting acceptance) may require evaluation under consumer contract notice requirements in certain jurisdictions including the EU and UK.